claude-mem
Claude-Mem is a JavaScript plugin that automatically captures and compresses context from AI agent sessions, storing relevant information and reinjecting it into future sessions. It works with Claude Code, OpenCode, and other AI agents to maintain project continuity across disconnected work periods.
Key facts
Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.
| Field | Value |
|---|---|
| Repository | thedotmack/claude-mem |
| Owner | thedotmack |
| Primary language | JavaScript |
| License | Apache-2.0 — OSI-approved |
| Stars | 86.4k |
| Forks | 7.5k |
| Open issues | 235 |
| Latest release | v13.10.2 (2026-07-05) |
| Last updated | 2026-07-08 |
| Source | https://github.com/thedotmack/claude-mem |
What claude-mem is
Node.js/JavaScript-based memory system using embeddings and semantic compression to capture tool observations, generate summaries, and perform retrieval-augmented generation (RAG) via SQLite and ChromaDB backends. Operates through lifecycle hooks in Claude Code and integrates with the OpenClaw gateway as a persistent worker service.
Get the claude-mem source
Clone the repository and explore it locally.
git clone https://github.com/thedotmack/claude-mem.gitcd claude-mem# follow the project's README for install & configurationNeed it deployed, integrated, or customized instead? DEV.co ships production installs.
Best use cases
Implementation considerations
- Requires Node.js ≥20.0.0; install via `npx claude-mem install` to properly register hooks and worker service, not via direct npm install.
- Memory storage backends (SQLite, ChromaDB) run locally; plan for disk space and database maintenance as memory grows over time.
- Embedding model selection affects quality and cost; configuration needed for AI provider (Anthropic, OpenAI, etc.) and token budgets.
- Use `<private>` tags in agent outputs to exclude sensitive data from storage; no automatic PII detection or redaction.
- Worker service runs as persistent Bun process; requires process management or systemd integration for production uptime.
When to avoid it — and what to weigh
- Single-session, stateless workflows — If your agent work is one-off or sessions are complete and discarded, the overhead of memory capture and retrieval will not provide value.
- Strict data residency or air-gapped environments — Claude-Mem stores compressed summaries locally but requires node runtime and external embedding models; offline-only scenarios or regulated data isolation may be problematic.
- Real-time, latency-critical agent responses — Memory retrieval and context injection add processing overhead; agents requiring sub-second response times may see noticeable latency during memory lookups.
- Teams with no Node.js or JavaScript infrastructure — Installation and configuration require npm/Node.js tooling and familiarity with Claude Code plugin architecture or OpenClaw gateways.
License & commercial use
Apache License 2.0 (Apache-2.0). Permissive OSI-approved open-source license allowing commercial use, modification, and distribution with proper attribution and liability/warranty disclaimers.
Apache-2.0 permits commercial use without explicit permission requirement. However, verify compatibility with your deployment environment (e.g., OpenClaw, embedding providers, data hosting) and review any third-party dependencies for their own license restrictions. No warranty or support guarantees in license text.
DEV.co evaluation signals
Editorial assessment — not user reviews. Directional, with an explicit confidence level.
| Signal | Assessment |
|---|---|
| Maintenance | Active |
| Documentation | Strong |
| License clarity | Clear |
| Deployment complexity | Moderate |
| DEV.co fit | Good |
| Assessment confidence | High |
Local storage of summaries and embeddings via SQLite/ChromaDB; no end-to-end encryption mentioned. <private> tag mechanism allows opt-out of memory capture but is manual and depends on agent discipline. No audit logging, access control, or data deletion workflow documented. Review data residency requirements before storing compressed agent context; verify embedding provider's data handling if using external models.
Alternatives to consider
Mem0
Standalone memory management platform for AI agents; offers managed SaaS and on-premise options with multi-agent support, but requires API integration rather than plugin-based deployment.
LangChain memory modules (ConversationBufferMemory, etc.)
Lightweight in-process memory primitives for Python/JS; simpler for development but lack cross-session persistence and semantic compression out of the box.
RAG frameworks (LlamaIndex, Pinecone vector store)
General-purpose retrieval-augmented generation systems; more flexible for custom backends and scaling but require more plumbing to integrate with agent workflows.
Build on claude-mem with DEV.co software developers
Start with `npx claude-mem install` to enable automatic context preservation across sessions. Review the docs for configuration, integrations, and best practices—or consult our team for production deployments.
Talk to DEV.coRelated on DEV.co
Explore the category and the services that help you build with it.
claude-mem FAQ
Does Claude-Mem work with local/self-hosted LLMs?
Can I export or backup stored memories?
What happens to memory if I switch embedding models?
How much does the worker service cost to run?
Work with a software development agency
From first prototype to production, DEV.co delivers software development services around tools like claude-mem. Our software development agency staffs experienced software developers and web developers for custom software development, web development, integrations, and ongoing support across rag frameworks and beyond.
Ready to Add Persistent Memory to Your AI Agents?
Start with `npx claude-mem install` to enable automatic context preservation across sessions. Review the docs for configuration, integrations, and best practices—or consult our team for production deployments.